Foundations: the mental model
Model selection, context windows, tokens, and the mechanics that quietly affect quality and cost.
Specialized training · Anthropic's Claude
Get your team productive with Claude—on work that matters.
Hands-on training for engineering, operations, and leadership. Practicing engineers adapt the curriculum to your stack so the team leaves with working patterns, not generic prompt tips.
- Delivery
- Remote, onsite, hybrid
- Team size
- 1-on-1 to 50 people
- Curriculum
- Custom to your stack
- Instruction
- Practicing engineers
Role-specific track
Engineering
For software teams integrating Claude into products, internal tools, and developer workflows. Hands-on with the APIs, SDKs, and code your team actually ships.
Tracks can stand alone or be combined. We remove modules that are irrelevant and add work specific to your environment.
Prompt patterns that survive production
System prompts, structured output, few-shot patterns, evaluation, and knowing when fine-tuning is the wrong answer.
Tool use, function calling, and agents
Tool design, agent loops, handoff rules, orchestration boundaries, and what should never be automated.
Claude Code as a daily driver
Skills, hooks, MCP servers, repository guidance, and development workflows that save real engineering time.
Cost, latency, and observability
Caching, batch requests, streaming, model routing, useful telemetry, and controlling surprise spend.
Hardening and failure modes
Prompt injection, retries, fallbacks, model migrations, and graceful degradation when a model misbehaves.
Role-specific track
Operations
For operations, customer success, and internal-tools teams adopting AI-assisted workflows. Less code; more emphasis on durable work inside existing processes.
Where Claude fits in existing tools
Practical patterns for Slack, Teams, Notion, internal portals, and other places where work already happens.
Document workflows
Reliable extraction, summarization, classification, and routing for contracts, invoices, reports, and email.
Customer-facing AI
PII handling, escalation paths, conversational guardrails, and deciding when Claude answers versus drafts.
Cost governance for non-engineers
Reading usage reports, setting alerts, and knowing when a workflow needs engineering or a different model.
Role-specific track
Leadership
For executives, founders, and team leads building an AI strategy. What is real, what is hype, and where to place the next practical bet.
What is actually shippable today
A grounded view of production-ready capabilities, rough edges, and demos that will not survive real users.
Build vs. buy vs. wait
A framework for choosing which capabilities to build, buy, or deliberately postpone.
Team enablement
Training paths by role, tool selection, operating habits, and moving AI from curiosity to repeatable capability.
Risk and lock-in
Prompt injection, data leakage, model drift, vendor concentration, and where audit checklists miss the actual exposure.
Formats
The shape that fits your team.
A focused starter
Half-day intensive
Foundations plus one track. Best for teams deciding what is real and what to try next.
Up to 25 people · Remote or onsite
Hands-on with your stack
Full-day workshop
Foundations plus two tracks, with lab time in your codebase or operational workflow.
Up to 20 people · Remote or onsite
Build something real together
Two-to-five-day deep dive
All three tracks plus a guided build of a real feature or workflow your team owns.
Up to 15 people · Onsite preferred
Embedded for the rollout
Ongoing coaching
Weekly or bi-weekly office hours, design reviews, and support for difficult implementation decisions.
3, 6, or 12 months · Remote
Outcomes
The Monday-after test.
Every engagement is designed around what your team can do differently when the session is over. We do not sell certificates.
A shared mental model
Your team can discuss Claude, its constraints, and its place in the stack using the same language.
A working starter
A codebase or workflow built in your environment during the engagement, ready for your team to extend.
A relevant pattern library
Prompts, tool definitions, and agent shapes grounded in your domain rather than tutorial examples.
A decision framework
Clear boundaries for what to automate, what stays human, and what should be revisited later.
A line back to the practitioners
A way to get answers when the difficult questions appear during implementation.
Why 159 Networks
We build Claude-backed systems and use Claude in our own engineering work. The material comes from real decisions about tool use, security, cost, failure modes, and maintainability—not an affiliated vendor course or a generic slide deck.
Questions
Before you book the curriculum call.
How long does an engagement take?
Anywhere from a half-day starter to a 12-month embedded engagement. Most teams begin with a two-to-three-day workshop and add coaching if useful.
What does it cost?
We quote per engagement rather than per seat. Scope, team size, format, and whether we work in your environment determine the price; the curriculum call is free.
Can you cover GPT, Gemini, or open models?
Most of the production patterns transfer across frontier models. Claude is our deepest day-to-day experience, but we can adapt the curriculum to another stack.
Are you affiliated with Anthropic?
No. We are an independent engineering shop that ships Claude-backed systems. We have no commercial incentive to recommend a tool that does not fit.
Can training be onsite?
Yes. US-based onsite delivery is standard, international work is quoted separately, and multi-day formats often benefit from being in the room together.
What if the team is brand new to AI?
That is a good fit. Foundations starts with the required mental model, and the Operations and Leadership tracks do not require a coding background.
Next step
Design the right curriculum for your team.
Tell us what your people are trying to do, and we will tell you whether training is the right move and what shape would produce a useful outcome.